Fine QTLs and genomic selection for production traits in ... · 11(+4) sires. G2 ewes. x. 11 Sarda...

Research Unit:

Genetics and Biotechnology

38th ICAR Annual Meeting. 2738th ICAR Annual Meeting. 27thth May May ‐‐ 11stst June 2012, Cork, IrelandJune 2012, Cork, Ireland

Agricultural Research Agency of Sardinia

A. Carta1, M.G. Usai1, S. Sechi1, S. Salaris1, S. Miari1, T. Sechi1, G.B. Congiu1, G. Mulas1, S. Murru2

and S. Casu1

1Research Unit: Genetics and Biotechnology, DIRPA‐AGRIS Sardegna, Sassari, Italy

2Associazione Nazionale della Pastorizia –

ASSONAPA, Rome, Italy

Fine mapping of QTLs and genomic selection for production traits in an experimental population of Sarda

dairy sheep.

Research Unit:




Official recording:•

Milk Yield• Fat and Protein Content• Udder Morphology• Somatic Cell Count• PrP Genotyping

Official recording:• Milk Yield• Fat and Protein Content• Udder Morphology• Somatic Cell Count• PrP Genotyping AI and NM

Elite rams

240,000 animals, 1,062 flocks

Sire’s dams

500 sampling rams ± 50 rams

3,000,000 animals; 11,000 flocks

One out‐of‐season lambing period per year

Lambing: adults from October to DecemberLambing: primiparous ewes from February to March

Lambs slaughtered after 30 days

Milk processed in PDO cheeses(Pecorino Romano, Pecorino Sardo e Fiore Sardo)

Diffusion of

genetic gain by

natural mating

rams

Selection scheme of Sarda dairy sheep breed

The breeding programs currently ongoing based on the traditional quantitative approach have achieved appreciable genetic gains for milk yield.

Introduction

Selected Population

Commercial Flocks



Official recording:• Milk Yield• Fat and Protein Content• Udder Morphology• Somatic Cell Count• PrP Genotyping

Research Unit:




Possibility of including other traits:

•SCC, FC, PC udder morphology (included with simplified recording)

•resistance to diseases (mastitis, internal parasites, paratuberculosis)

•milk nutritional value (fatty acid composition)

•milkability traits

limited by

high organizational effort

needed to apply the traditional quantitative approach

high recording costs for traits difficult to measure

reduction of public funding

The application of selection schemes assisted by molecular

information is potentially useful in dairy sheep

Research Unit:




The main reasons:

•high cost of HD SNP arrays mainly if related to the number of recorded traits

•difficulty in finding well‐structured training populations to estimate SNP effects

• advent of affordable high‐throughput technology for SNP

• reduction in sequencing costs

shift to SNP markers for

QTL mapping and genome‐wide selection studies

Genome‐wide selection seems still unachievable

in most dairy sheep breeds

Research Unit:




To present preliminary elements to evaluate the

potential evolution to nucleus flocks of pre‐existing

experimental populations created for QTL detection

purposes

to increase the efficiency of the selection process

Aim of this study

Experimental populationx

10 Lacaune sires 10 Sarda dams

10 F1 siresLacaune chromosomes=

Sarda chromosomes=

1998

Sarda chromosomes=

xSarda ewes

BC ewes (G0)

2000‐2004

G3 ewes

x18 Sarda sires

G1 ewesx

11(+4) sires

G2 ewesx

11 Sarda sires

2003‐2009

x~20 AI sires/year

G3/G4 ewes

2010…

Research Unit:




S2

has been designed to be the

nucleus flock (NF) of the

Sarda breed with the aim of estimating EBV and DGV of

the most important blood lines

Materials and methods

Description of the experimental nucleus flock of the Sarda breed

BC

generated by mating F1 Lacaune x Sarda sires with Sarda ewesS1

generated by mating Sarda sires with S0 ewesS2

generated by mating AI Sarda rams with S1 ewes

Research Unit:




PhenotypesMaterials and methods

Research Unit:




Genotyped animals (Illumina Inc. OvineSNP50 Beadchip):

• Around 2,000 G0, G1

and G2 ewes

• 10 F1 Lacaune x Sarda sires,

• 62 out of the 88 Sarda sires used in nucleus flock (SA)

• 94 AI old

Sarda rams with the highest genetic impact

on the selected population (HI)

Genotypes Genotyping: Porto Conte Ricerche ‐

iScan Platform of Illumina.

Individuals with a portion of missing genotypes higher than 10%

X chromosome and SNPs which could not be mapped

SNPs with call rate lower than 95%

MAF lower than 1%

(final data set: 44,859 SNP)


Research Unit:




• LA with “extended”

families:

including phenotypes and genotypes of grand‐

and great‐grand‐daughters

• At each SNP position a within‐family linear regression of phenotypes on the probability of inheriting one of the QTL

alleles of the family founder

• The model also included a founder effect and was tested by likelihood ratio test (LRT).

• Separate analyses were performed for F1

and SA

families in order to detect QTL segregating between breeds and within the Sarda breed respectively.

• Chromosome‐wise and genome‐wise significance thresholds were defined by 10,000 within family permutations.

Fine mapping of QTLs


Research Unit:




• Evaluated by relationship

and genomic relationship matrices

• Relationship matrix included 2,894 selected population (SP) sires of ewes with lactation records in 2011 and SA rams.

• The genomic relationship matrix included 94 HI and 62 SA

rams.

Genetic link between the selected population and the nucleus flock


Research Unit:




DGV were estimated

by LASSO‐LARS (LL) procedure using

1,464 ewes

of G1 and G2.

LL

procedure was run until 500 SNPs were fitted in the model.

At each step the correlation between DGVs obtained by the current set of active SNP effects and the corresponding EBVs was calculated.

The DGV for MY of HI and SA rams was calculated as the sum of the genotype effects derived from the estimated SNP allele substitution effects.

DGV and EBV calculation


Research Unit:




1.

SA rams: within NF and

official EBV

s

2.

SA rams: EBV

estimated on NF daughters and DGV

: to estimate the

SNP effects’

predictive ability

3.

HI rams: EBV estimated in the official genetic evaluation and DGV

:

to estimate the SNP effects’

predictive ability of EBV estimated for SP

rams

DGV and EBV correlations


Research Unit:




Results and discussion

Milk Yield• 2 QTL exceeding the genome‐wise significance threshold in F1 families• No SA QTL was genome‐wise significant.

Fat Content• 5 QTL exceeded the 0.05 genome‐wise significance threshold in F1 families. • 2 QTL exceeded the 0.05 genome‐wise significance in SA sires

Protein Content• 4 QTL exceeding the genome‐wise significance threshold in F1 families• 1 QTL exceeding the genome‐wise significant threshold in SA.

. * chromosome‐wise significant (p<0.05); § genome‐wise significant (p<0.05)

QTL fine mapping

Number of QTL detected in F1 was larger than in SA sires

Research Unit:




• Among the 16 QTLs in F1 and SA on the same chromosomes, only OAR20 for MY, OAR 6 and 16 for PC

showed close peaks (<5 Mb)

• Results do not allow to clearly define causative mutation or LD markers

• LD and LDLA analyses are expected to improve the power of fine mapping

• Further improvements are expected by the increasing of the size of the genotyped and phenotyped portion of NF

and by the sequencing of a set of

target animals

. * chromosome‐wise significant (p<0.05); § genome‐wise significant (p<0.05)QTL fine mapping


Research Unit:




• 90% of HI have a genomic relationship higher than 0.125 with at least one SA sire

• 80% of SP sires had a relationship coefficient higher than 0.0625 with at least one SA.

• The percentage of HI and SP genome represented in SA seems to be sufficient to allow either a correct estimation of SNP effects or a sufficiently accurate evaluation

of SP blood lines.

Frequencies of maximum relationship between pairs of SP and HI sires with SA sires extracted from the

relationship and genomic relationship matrix respectively.

Genetic link between selected population and nucleus flock


Research Unit:




• Correlation between SA within NF EBV and official EBV was 0.51

• Correlation between SA within NF EBV and

DGV

was 0.91

• Correlation between official EBV and DGV of HI sires was 0.43.

DGV and EBV estimation

These results confirm that SNP effects estimated by genotypes of NF ewe

have a promising predictive ability of SA and HI sires EBVs.

The genomic tools originally used to detect QTLs can be also used also for genome‐wide selection

Further improvements are expected by increasing the number of genotyped and phenotyped NF ewes.


Research Unit:




Conclusion

QTL fine mapping•The

addition

of

the

following

generations

of

genotyped

and

phenotyped

ewes

combined

with

more

sophisticated

statistical methods will lead to a more precise definition of the QTL regions.

High

density

sequencing

techniques

on

target

animals

should

allow to identify causal mutations or LD SNPs to be used for selection.

DGV and EBV estimation•The

correlation

between

the

DGV

based

on

SNP

effects

from

the

nucleus

flock

and

EBV

estimated

by

progeny

test

in

the

selected population

confirms

that

the

nucleus

flock

can

be

useful

to

predict

genetic merit of rams used in the selected population.

Research Unit:




• Results seem promising for including the pre‐existing experimental flock

created

for

QTL

detection

purposes

in

the

breeding

program

of the Sarda breed with

the aim

of

increasing

the

efficiency

of

the selection process.

• In the next future, simulation studies will be carried out to optimize the

size

of

the

nucleus

flock

and

to

find

objective

methods

for

choosing

the

blood

lines

to

include

in

the

nucleus

flock

and defining the size of the sire families.

General Conclusions

Research Unit:




Potential Evolution of selection scheme of Sarda dairy sheep brePotential Evolution of selection scheme of Sarda dairy sheep breeded

AI and NM Elite rams

240,000 animals, 1,062 flocks

Sire’s dams

500 sampling rams

± 50 rams

3,000,000 animals; 11,000 flocks

One out‐of‐season lambing period per year

Lambing: adults from October to DecemberLambing: primiparous ewes from February to March

Lambs slaughtered after 30 days

Milk processed in PDO cheeses(Pecorino Romano, Pecorino Sardo e Fiore Sardo)

Selected Population

Commercial Flocks




Official recording:• Milk Yield• Fat and Protein Content• Udder Morphology• Somatic Cell Count• PrP Genotyping

Research Unit:




Thank you for your attention

Study realised in the framework of the European Project 3SR ( FP7‐KBBE‐2009‐3‐245140)

Fine QTLs and genomic selection for production traits in ... · 11(+4) sires. G2 ewes. x. 11 Sarda...

Documents

Transcript of Fine QTLs and genomic selection for production traits in ... · 11(+4) sires. G2 ewes. x. 11 Sarda...